//--------------------------------------------------------------------------------// // COPYRIGHT NOTICE // //--------------------------------------------------------------------------------// // Copyright (c) 2012, Instituto de Microelectronica de Sevilla (IMSE-CNM) // // // // All rights reserved. // // // // Redistribution and use in source and binary forms, with or without // // modification, are permitted provided that the following conditions are met: // // // // * Redistributions of source code must retain the above copyright notice, // // this list of conditions and the following disclaimer. // // // // * Redistributions in binary form must reproduce the above copyright // // notice, this list of conditions and the following disclaimer in the // // documentation and/or other materials provided with the distribution. // // // // * Neither the name of the IMSE-CNM nor the names of its contributors may // // be used to endorse or promote products derived from this software // // without specific prior written permission. // // // // THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" // // AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE // // IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE // // DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE // // FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL // // DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR // // SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER // // CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, // // OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE // // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. // //--------------------------------------------------------------------------------// //++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++// // ALGORITMO DE DESCENSO POR SIMPLEX // //++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++// package xfuzzy.xfsl.algorithm; import xfuzzy.xfsl.*; import xfuzzy.lang.*; import java.util.Random; public class Simplex extends XfslAlgorithm { //+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++// // MIEMBROS PRIVADOS // //+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++// private double pert; private double inc; private double dec; private double[][] simplex; private double[] simsum; private XfslEvaluation[] simeval; private int ilo, ihi, inhi; private double lasterror; private Random random; //+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++// // CONSTRUCTOR // //+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++// public Simplex() { this.pert = -1; this.inc = -1; this.dec = -1; this.random = new Random(); } //+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++// // METODOS PUBLICOS // //+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++// //-------------------------------------------------------------// // Devuelve el codigo de identificacion del algoritmo // //-------------------------------------------------------------// public int getCode() { return SIMPLEX; } //-------------------------------------------------------------// // Actualiza los parametros de configuracion del algoritmo // //-------------------------------------------------------------// public void setParameters(double[] param) throws XflException { if(param.length != 3) throw new XflException(26); pert = test(param[0], POSITIVE); inc = test(param[1], INCREASE); dec = test(param[2], DECREASE); } //-------------------------------------------------------------// // Obtiene los parametros de configuracion del algoritmo // //-------------------------------------------------------------// public XfslAlgorithmParam[] getParams() { XfslAlgorithmParam[] pp = new XfslAlgorithmParam[3]; pp[0] = new XfslAlgorithmParam(pert, POSITIVE, "Perturbation"); pp[1] = new XfslAlgorithmParam(inc, INCREASE, "Expansion Factor"); pp[2] = new XfslAlgorithmParam(dec, DECREASE, "Contraction Factor"); return pp; } //-------------------------------------------------------------// // Obtiene las opciones de configuracion del algoritmo // //-------------------------------------------------------------// public XfslAlgorithmOption[] getOptions() { XfslAlgorithmOption[] opt = new XfslAlgorithmOption[0]; return opt; } //-------------------------------------------------------------// // Ejecuta una iteracion del algoritmo // //-------------------------------------------------------------// public XfslEvaluation iteration(Specification spec, XfslPattern pattern, XfslErrorFunction ef) throws XflException { Parameter[] param = spec.getAdjustable(); if(init) { init = false; simplex = new double[param.length+1][param.length]; simsum = new double[param.length]; simeval = new XfslEvaluation[param.length+1]; for(int j=0; j<param.length; j++) simplex[0][j] = param[j].value; for(int j=0; j<param.length; j++) simsum[j] += param[j].value; simeval[0] = ef.evaluate(spec,pattern,1.0); lasterror = simeval[0].error; for(int i=1; i<simplex.length; i++) { perturb(spec); for(int j=0; j<param.length; j++) simplex[i][j] = param[j].value; for(int j=0; j<param.length; j++) simsum[j] += param[j].value; simeval[i] = ef.evaluate(spec,pattern,lasterror); for(int j=0; j<param.length; j++) param[j].value = simplex[0][j]; } getLimits(); for(int j=0; j<param.length; j++) param[j].value = simplex[ilo][j]; lasterror = simeval[ilo].error; return simeval[ilo]; } XfslEvaluation simtry = amotry(spec,pattern,ef,param,ihi,-1.0); if(simtry.error <= simeval[ilo].error) simtry = amotry(spec,pattern,ef,param,ihi,inc); else if(simtry.error >= simeval[inhi].error) { XfslEvaluation simsave = simeval[ihi]; simtry = amotry(spec,pattern,ef,param,ihi,dec); if(simtry.error >= simsave.error) { for(int i=0; i<simplex.length; i++) if(i != ilo) { for(int j=0; j<param.length; j++) param[j].setDesp(dec*(simplex[i][j] - simplex[ilo][j])); spec.update(); simeval[i] = ef.evaluate(spec,pattern,1.0); for(int j=0; j<param.length; j++) simplex[i][j] = param[j].value; for(int j=0; j<param.length; j++) param[j].value = simplex[ilo][j]; } for(int j=0; j<param.length; j++) { simsum[j] = 0; for(int i=0; i<simplex.length; i++) simsum[j] += simplex[i][j]; } } } getLimits(); for(int j=0; j<param.length; j++) param[j].value = simplex[ilo][j]; simeval[ilo].var = (lasterror - simeval[ilo].error)/lasterror; lasterror = simeval[ilo].error; return simeval[ilo]; } //+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++// // METODOS PRIVADOS // //+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++// //-------------------------------------------------------------// // Evalua una proyeccion de factor fac // //-------------------------------------------------------------// private XfslEvaluation amotry(Specification spec, XfslPattern pattern, XfslErrorFunction ef, Parameter[] param, int ihi, double fac) { double[] ptry = new double[ simplex[ihi].length ]; double fac1 = (1.0 - fac)/simplex[ihi].length; double fac2 = fac1 - fac; for(int i=0; i<param.length; i++) { param[i].value = simplex[ihi][i]; ptry[i] = simsum[i]*fac1 - simplex[ihi][i]*fac2; param[i].setDesp(ptry[i] - simplex[ihi][i]); } spec.update(); XfslEvaluation ytry = ef.evaluate(spec,pattern,1.0); if(ytry.error < simeval[ihi].error) { simeval[ihi] = ytry; for(int i=0; i<param.length; i++) { simsum[i] += param[i].value - simplex[ihi][i]; simplex[ihi][i] = param[i].value; } } for(int i=0; i<param.length; i++) param[i].value = simplex[ilo][i]; return ytry; } //-------------------------------------------------------------// // Actualiza los indices del mejor, peor y antepeor // //-------------------------------------------------------------// private void getLimits() { ilo = 0; ihi = (simeval[0].error > simeval[1].error ? 0 : 1); inhi = (simeval[0].error > simeval[1].error ? 1 : 0); for(int i=0; i<simplex.length; i++) { if(simeval[i].error <= simeval[ilo].error) ilo = i; if(simeval[i].error > simeval[ihi].error) { inhi = ihi; ihi = i; } else if(simeval[i].error > simeval[inhi].error && i!=ihi) inhi = i; } } //-------------------------------------------------------------// // Realiza una perturbacion del sistema // //-------------------------------------------------------------// private void perturb(Specification spec) { Type[] type = spec.getTypes(); for(int i=0; i<type.length; i++) { if(!type[i].isAdjustable()) continue; Family fam[] = type[i].getFamilies(); for(int j=0; j<fam.length; j++) { if(!fam[j].isAdjustable()) continue; Parameter[] pp = fam[j].getParameters(); double[] val = fam[j].get(); do { for(int k=0; k<pp.length; k++) { if(!pp[k].isAdjustable()) continue; pp[k].value = val[k] + pert*(random.nextDouble()-0.5); } } while(!fam[j].test()); } ParamMemFunc[] mf = type[i].getParamMembershipFunctions(); for(int j=0; j<mf.length; j++) { if(!mf[j].isAdjustable()) continue; Parameter[] pp = mf[j].getParameters(); double[] val = mf[j].get(); do { for(int k=0; k<pp.length; k++) { if(!pp[k].isAdjustable()) continue; pp[k].value = val[k] + pert*(random.nextDouble()-0.5); } } while(!mf[j].test()); } } } }